BACKGROUND: Low Chlamydia trachomatis screening rates create an opportunity to test innovative continuing medical education (CME) programs. Few studies of Internet-based physician learning have been evaluated with objective data on practice patterns. DESIGN: This randomized controlled trial tested a multicomponent Internet CME (mCME) intervention for increasing chlamydia screening of at-risk women aged 16 to 26 years. SETTING: Eligible physician offices had > or =20 patients at risk for chlamydia as defined by the Health Plan Employer Data and Information Set (HEDIS), had at least one primary care physician (internal medicine, family medicine/general practice, pediatrics) with Internet access, and participated in the study managed care organization. The 191 randomized primary care offices represented 20 states. INTERVENTION: The intervention, available from February to December 2001, consisted of four case-based learning modules, was tailored in real time to each physician based on theory of behavior change, and included office-level feedback of chlamydia screening rates. MAIN OUTCOME MEASURE: HEDIS chlamydia screening rates for the pre-intervention (2000) and post-intervention (2002) periods. RESULTS: Pre-intervention screening rates for the intervention and comparison offices were 18.9% and 16.2% (p =0.135). Post-intervention screening rates for the intervention and comparison offices were 15.5% and 12.4%, respectively (p =0.044, adjusting for baseline performance). CONCLUSIONS: The substantial decline in chlamydia screening rates observed in the comparison offices was significantly attenuated for the intervention offices. The mCME favorably influenced chlamydia screening by primary care physicians.